5 A C Compiler for 6502 Systems
7 (C) Copyright 1989 John R. Dunning
8 (C) Copyright 1998-2000 Ullrich von Bassewitz
22 4. Differences to the ISO standard
39 cc65 was originally a C compiler for the Atari 8-bit machines written by
40 John R. Dunning. In prior releases I've described the compiler by listing
41 up the changes made by me. I have made many more changes in the meantime
42 (and rewritten major parts of the compiler), so I will no longer do that,
43 since the list would be too large and of no use to anyone. Instead I will
44 describe the compiler in respect to the ANSI/ISO C standard. In fact, I'm
45 planning a complete rewrite (that is, a complete new compiler) for the
46 next release, since there are too many limitations in the current code,
47 and removing these limitations would mean a rewrite of many more parts of
50 There is a separate document named "library.txt" that covers the library
51 available for the compiler. If you know C and are interested in doing
52 actual programming, the library documentation is probably of much more use
55 If you need some hints for getting the best code out of the compiler, you
56 may have a look at "coding.txt" which covers some code generation issues.
63 The compiler translates C files into files containing assembler code that
64 may be translated by the ca65 macroassembler (for more information about
65 the assembler, have a look at ca65.txt).
67 The compiler may be called as follows:
69 Usage: cc65 [options] file
71 -g Add debug info to object files
73 -j Default characters are signed
74 -o name Name the output file
75 -s Print some statistics
76 -tx Set target system x
79 -Cl Make local variables static
80 -Dsym[=defn] Define a symbol
81 -I path Set include directory
83 -Oi Optimize code, inline more code
84 -Or Enable register variables
85 -Os Inline some known functions
86 -T Include source as comment
87 -V Print version number
90 The -A option disables any compiler exensions. Have a look at section 5
91 for a discussion of compiler extensions. In addition, the macro
95 is defined, when compiling with -A.
97 -d enables debug mode, something that should not be needed for mere
100 -g will cause the compiler to insert a .DEBUGINFO command into the
101 generated assembler code. This will cause the assembler to include all
102 symbols in a special section in the object file.
104 -h and -s print some statistics, nothing spectacular.
106 Using -j you can make the default characters signed. Since the 6502 has
107 no provisions for sign extending characters (which is needed on almost
108 any load operation), this will make the code larger and slower. A better
109 way is to declare characters explicitly as "signed" if needed. You can
110 also use "#pragma signedchars" for better control of this option (see
113 The -t option is used to set the target system. The target system
114 determines things like the character set that is used for strings and
115 character constants. The following target systems are supported:
120 ace (no library support)
123 pet (all CBM PET systems except the 2001)
124 nes (Nintendo Entertainment System)
128 Using -v, the compiler will be somewhat more verbose if errors or warnings
131 -Cl will use static storage for local variables instead of storage on the
132 stack. Since the stack is emulated in software, this gives shorter and
133 usually faster code, but the code is no longer reentrant. The difference
134 between -Cl and declaring local variables as static yourself is, that
135 initializer code is executed each time, the function is entered. So when
144 the variable a will always have the value 1 when entering the function and
149 static unsigned a = 1;
153 the variable a will have the value 1 only the first time, the function is
154 entered, and will keep the old value from one call of the function to the
157 You may also use #pragma staticlocals to change this setting in your
158 sources (see section 7).
160 -I sets the directory where the compiler searches for include files. You
161 may use -I multiple times to add more than one directory to the search
164 -O will enable an optimizer run over the produced code. Using -Oi, the
165 code generator will inline some code where otherwise a runtime functions
166 would have been called, even if the generated code is larger. This will
167 not only remove the overhead for a function call, but will make the code
168 visible for the optimizer.
170 -Or will make the compiler honor the "register" keyword. Local variables
171 may be placed in registers (which are actually zero page locations).
172 There is some overhead involved with register variables, since the old
173 contents of the registers must be saved and restored. In addition, the
174 current implementation does not make good use of register variables, so
175 using -Or may make your program even slower and larger. Use with care!
177 Using -Os will force the compiler to inline some known functions from the
178 C library like strlen. Note: This has two consequences:
180 * You may not use names of standard C functions in your own code. If
181 you do that, your program is not standard compliant anyway, but
182 using -Os will actually break things.
184 * The inlined string and memory functions will not handle strings or
185 memory areas larger than 255 bytes. Similar, the inlined is..()
186 functions will not work with values outside char range.
188 It is possible to concatenate the modifiers for -O. For example, to
189 enable register variables and inlining of known functions, you may use
192 -T will include the source code as comments in the generated code. This is
195 -V prints the version number of the compiler. When submitting a bug
196 report, please include the operating system you're using, and the compiler
199 The -W switch suppresses any warnings generated by the compiler. Since any
200 source file may be written in a manner that it will not produce compiler
201 warnings, using this option is usually not a good idea.
208 The compiler will accept one C file per invocation and create a file with
209 the same base name, but with the extension replaced by ".s". The output
210 file contains assembler code suitable for the use with the ca65 macro
213 In addition to the paths named in the -I option on the command line, the
214 directory named in the environment variable CC65_INC is added to the
215 search path for include files on startup.
219 4. Differences to the ISO standard
220 ----------------------------------
222 Here is a list of differences between the language, the compiler accepts,
223 and the one defined by the ISO standard:
226 * The compiler allows single line comments that start with //. This
227 feature is disabled in strict ANSI mode.
229 * The compiler allows unnamed parameters in parameter lists. The
230 compiler will not issue warnings about unused parameters that don't
231 have a name. This feature is disabled in strict ANSI mode.
233 * The compiler has some additional keywords:
235 asm, __asm__, fastcall, __fastcall__, __AX__, __EAX__, __func__
237 The keywords without the underlines are disabled in strict ANSI mode.
239 * The "const" modifier is available, but has no effect.
241 * The datatypes "float" and "double" are not available.
243 * The compiler does not support bit fields.
245 * Initialization of local variables is only possible for scalar data
246 types (that is, not for arrays and structs).
248 * Because of the "wrong" order of the parameters on the stack, there is
249 an additional macro needed to access parameters in a variable
250 parameter list in a C function.
252 * The compiler has only one symbol table. Because of that, it's not
253 possible to use the name of a local variable in a nested block in the
254 same function (global and local names are distinct, however).
256 + The preprocessor does not understand the "defined" keyword in
257 expressions evaluated in #if statements.
259 * Functions may not return structs, struct assignment is not possible.
261 * The size of any struct referenced via a pointer may not exceed 256
262 bytes (this is because the Y register is used as index).
264 * In a function, the size of the parameters plus the size of all local
265 variables may not exceed 256 bytes (in fact, the limit may be even less
266 depeding on the complexity of your expressions).
268 * Part of the C library is available only with fastcall calling
269 conventions (see below). This means, that you may not mix pointers to
270 those functions with pointers to user written functions.
272 There may be some more minor differences, I'm currently not aware off. The
273 biggest problems are the missing const and float data types. With both
274 these things in mind, you should be able to write fairly portable code.
281 This cc65 version has some extensions to the ISO C standard.
283 * The compiler allows // comments (like in C++ and in the proposed C9x
284 standard). This feature is disabled by -A.
286 * The compiler allows to insert assembler statements into the output
289 asm (<string literal>) ;
293 __asm__ (<string literal>) ;
295 The first form is in the user namespace and is disabled if the -A
298 The given string is inserted literally into the output file, and a
299 newline is appended. The statements in this string are not checked by
300 the compiler, so be careful!
302 The asm statement may be used inside a function and on global file
305 * There is a special calling convention named "fastcall". This calling
306 convention is currently only usable for functions written in
307 assembler. The syntax for a function declaration using fastcall is
309 <return type> fastcall <function name> (<parameter list>)
313 <return type> __fastcall__ <function name> (<parameter list>)
317 void __fastcall__ f (unsigned char c)
319 The first form of the fastcall keyword is in the user namespace and is
320 therefore disabled in strict ANSI mode.
322 For functions declared as fastcall, the rightmost parameter is not
323 pushed on the stack but left in the primary register when the function
324 is called. This will reduce the cost when calling assembler functions
325 significantly, especially when the function itself is rather small.
327 BEWARE: You must not declare C functions as fastcall! This will not
328 work for now and is not checked by the assembler, so you will get
331 * There are two pseudo variables named __AX__ and __EAX__. Both refer to
332 the primary register that is used by the compiler to evaluate
333 expressions or return function results. __AX__ is of type unsigned int
334 and __EAX__ of type long unsigned int respectively. The pseudo
335 variables may be used as lvalue and rvalue as every other variable.
336 They are most useful together with short sequences of assembler code.
337 For example, the macro
339 #define hi(x) (__AX__=(x),asm("\ttxa\n\tldx\t#$00",__AX__)
341 will give the high byte of any unsigned value.
343 * Inside a function, the identifier __func__ gives the name of the
344 current function as a string. Outside of functions, __func__ is
348 #define PRINT_DEBUG(s) printf ("%s: %s\n", __func__, s);
350 The macro will print the name of the current function plus a given
358 The compiler defines several macros at startup:
361 __CC65__ This macro is always defined. Its value is the version
362 number of the compiler in hex. Version 2.0.1 of the
363 compiler will have this macro defined as 0x0201.
365 __CBM__ This macro is defined if the target system is one of the
368 __C64__ This macro is defined if the target is the c64 (-t c64).
370 __C128__ This macro is defined if the target is the c128 (-t c128).
372 __PLUS4__ This macro is defined if the target is the plus/4
375 __CBM610__ This macro is defined if the target is one of the CBM
376 600/700 family of computers (called B series in the US).
378 __PET__ This macro is defined if the target is the PET family of
381 __NES__ This macro is defined if the target is the Nintendo
382 Entertainment System (-t nes).
384 __ATARI__ This macro is defined if the target is one of the Atari
385 computers (400/800/130XL/800XL). Note that there is no
386 runtime and C library support for atari systems.
388 __ACE__ This macro is defined if the target is Bruce Craigs ACE
389 operating system. Note that there is no longer runtime
390 and library support for ACE.
392 __APPLE2__ This macro is defined if the target is the Apple ][
395 __GEOS__ This macro is defined if you are compiling for the GEOS
398 __FILE__ This macro expands to a string containing the name of
401 __LINE__ This macro expands to the current line number.
403 __STRICT_ANSI__ This macro is defined to 1 if the -A compiler option was
404 given, and undefined otherwise.
406 __OPT__ Is defined if the compiler was called with the -O command
409 __OPT_i__ Is defined if the compiler was called with the -Oi command
412 __OPT_r__ Is defined if the compiler was called with the -Or command
415 __OPT_s__ Is defined if the compiler was called with the -Os command
423 The compiler understands some pragmas that may be used to change code
424 generation and other stuff.
427 #pragma bssseg (<name>)
429 This pragma changes the name used for the BSS segment (the BSS segment
430 is used to store uninitialized data). The argument is a string enclosed
433 Note: The default linker configuration file does only map the standard
434 segments. If you use other segments, you have to create a new linker
437 Beware: The startup code will zero only the default BSS segment. If you
438 use another BSS segment, you have to do that yourself, otherwise
439 uninitialized variables do not have the value zero.
443 #pragma bssseg ("MyBSS")
446 #pragma codeseg (<name>)
448 This pragma changes the name used for the CODE segment (the CODE segment
449 is used to store executable code). The argument is a string enclosed in
452 Note: The default linker configuration file does only map the standard
453 segments. If you use other segments, you have to create a new linker
458 #pragma bssseg ("MyCODE")
461 #pragma dataseg (<name>)
463 This pragma changes the name used for the DATA segment (the DATA segment
464 is used to store initialized data). The argument is a string enclosed in
467 Note: The default linker configuration file does only map the standard
468 segments. If you use other segments, you have to create a new linker
473 #pragma bssseg ("MyDATA")
476 #pragma rodataseg (<name>)
478 This pragma changes the name used for the RODATA segment (the RODATA
479 segment is used to store readonly data). The argument is a string
480 enclosed in double quotes.
482 Note: The default linker configuration file does only map the standard
483 segments. If you use other segments, you have to create a new linker
488 #pragma bssseg ("MyRODATA")
491 #pragma regvaraddr (<const int>)
493 The compiler does not allow to take the address of register variables.
494 The regvaraddr pragma changes this. Taking the address of a register
495 variable is allowed after using this pragma, if the argument is not
496 zero. Using an argument of zero changes back to the default behaviour.
498 Beware: The C standard does not allow taking the address of a variable
499 declared as register. So your programs become non-portable if you use
500 this pragma. In addition, your program may not work. This is usually the
501 case if a subroutine is called with the address of a register variable,
502 and this subroutine (or a subroutine called from there) uses itself
503 register variables. So be careful with this #pragma.
507 #pragma regvaraddr(1) /* Allow taking the address
508 * of register variables
512 #pragma signedchars (<const int>)
514 Changed the signedness of the default character type. If the argument
515 is not zero, default characters are signed, otherwise characters are
516 unsigned. The compiler default is to make characters unsigned since this
517 creates a lot better code.
520 #pragma staticlocals (<const int>)
522 Use variables in the bss segment instead of variables on the stack. This
523 pragma changes the default set by the compiler option -Cl. If the argument
524 is not zero, local variables are allocated in the BSS segment, leading to
525 shorter and in most cases faster, but non-reentrant code.
528 #pragma zpsym (<name>)
530 Tell the compiler that the - previously as external declared - symbol with
531 the given name is a zero page symbol (usually from an assembler file).
532 The compiler will create a matching import declaration for the assembler.
537 #pragma zpsym ("foo"); /* foo is in the zeropage */
544 If you have problems using the compiler, if you find any bugs, or if
545 you're doing something interesting with the compiler, I would be glad to
546 hear from you. Feel free to contact me by email (uz@musoftware.de).
553 This is the original compiler copyright:
555 --------------------------------------------------------------------------
558 This is the copyright notice for RA65, LINK65, LIBR65, and other
559 Atari 8-bit programs. Said programs are Copyright 1989, by John R.
560 Dunning. All rights reserved, with the following exceptions:
562 Anyone may copy or redistribute these programs, provided that:
564 1: You don't charge anything for the copy. It is permissable to
565 charge a nominal fee for media, etc.
567 2: All source code and documentation for the programs is made
568 available as part of the distribution.
570 3: This copyright notice is preserved verbatim, and included in
573 You are allowed to modify these programs, and redistribute the
574 modified versions, provided that the modifications are clearly noted.
576 There is NO WARRANTY with this software, it comes as is, and is
577 distributed in the hope that it may be useful.
579 This copyright notice applies to any program which contains
580 this text, or the refers to this file.
582 This copyright notice is based on the one published by the Free
583 Software Foundation, sometimes known as the GNU project. The idea
584 is the same as theirs, ie the software is free, and is intended to
585 stay that way. Everybody has the right to copy, modify, and re-
586 distribute this software. Nobody has the right to prevent anyone
587 else from copying, modifying or redistributing it.
589 --------------------------------------------------------------------------
591 In acknowledgment of this copyright, I will place my own changes to the
592 compiler under the same copyright. Please note however, that the library
593 and all binutils are covered by another copyright, and that I'm planning
594 to do a complete rewrite of the compiler, after which the compiler
595 copyright will also change.
597 For the list of changes requested by this copyright see newvers.txt.